Complex and collaborative problem solving

Complex and collaborative problem solving names a pair of transversal skills for work where the problem structure is dynamic, information must be actively acquired, and useful performance depends on applying knowledge under uncertainty either alone or with others. 2015-neubert-assessment-21st-century-skills argues that CPS and ColPS are especially useful for I-O psychology because they sit between broad latent attributes such as intelligence and narrower practice labels such as learning agility, talent, or leadership potential.

The assessment contribution is the important part for this wiki: CPS and ColPS can be operationalized through computer-based microworlds that vary problem features systematically, simulate active exploration and control, and collect process data about how people search, coordinate, and adapt. That makes them a candidate bridge between job-analysis and construct-validity: close enough to nonroutine and interactive work to matter, but theoretically specified enough to avoid becoming another loose competency label.

For AI workforce transformation, CPS and ColPS are a reminder that “future skills” claims need assessment designs that represent changing work demands rather than merely rename broad capability aspirations. As AI absorbs routine production and changes the human role toward diagnosis, coordination, exception handling, and verification, the useful question is not whether people have a generic future skill; it is whether the assessment captures the problem features, collaboration demands, information search, and adaptive control that the redesigned work actually requires.

CPS and ColPS are capability constructs; critical-thinking is currently an AI-impact topic. The bridge is that AI-mediated work may shift which parts of complex and collaborative problem solving humans still perform: problem framing, information search, verification, coordination, exception handling, and adaptive control. The construct page asks how those skills should be defined and assessed; the critical-thinking page asks whether GenAI use preserves, relocates, overloads, or erodes judgment during those tasks.

2025-teyssier-roberge-21st-century-skills-overlap reinforces why CPS and ColPS should not be treated as just two more labels in a long future-skills inventory. The paper’s overlap analysis suggests that useful AI-era skill constructs need to be consolidated around separable higher-order competencies and then assessed through behavior-rich tasks, which is the reason CPS/ColPS remain promising only if their task features map to actual work demands.

Connections

• Relates to job-analysis because CPS and ColPS respond to the limits of task descriptions anchored in existing occupations when work becomes more nonroutine, dynamic, and collaborative.
• Relates to competency-modeling because broad skill labels become safer when they are translated into observable problem-solving behavior and assessment tasks rather than left as aspirational capability language.
• Relates to construct-validity because CPS and ColPS need clear nomological placement, systematic task-feature variation, evidence that assessment scores support the intended workplace inferences, and protection against the same semantic and psychometric overlap that weakens broad 21st-century-skills frameworks.
• Relates to human-capital-resource-measurement because collective capability claims depend partly on whether individual and collaborative problem-solving skills are measured as actual resources rather than credentials, experiences, or outcomes.
• Relates to work-redesign because AI-mediated roles may increase the relative importance of human problem framing, coordination, exception handling, and adaptive control even as routine task execution changes.
• Relates to critical-thinking because critical thinking is one judgment process that may be exercised within CPS/ColPS tasks, especially during verification, problem framing, and adaptive control under AI assistance.

Contradictions & open questions

• 2015-neubert-assessment-21st-century-skills treats CPS and ColPS as promising bridges between theory and practice, but most of the article’s assessment evidence comes from education and large-scale testing rather than direct organizational validation.
• A microworld can improve construct representation compared with questionnaires, but it can still miss accountable workplace context, incentives, professional norms, and the messy coordination costs of real work.
• The AI-era version of this problem is still open: which human problem-solving and collaboration demands become more important when AI systems handle more routine analysis, and which demands are reduced, delegated, or transformed?
• 2025-teyssier-roberge-21st-century-skills-overlap shows why “future skills” need pruning, but pruning alone does not solve the work-design problem: AI may change which problem features, collaboration demands, and adaptive-control behaviors are actually criterion-relevant.